Note: This copy is for your personal, non-commercial use only.
To order presentation-ready copies for
n GENITOURINARY IMAGING distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.
Bosniak Category IIF and III
Cystic Renal Lesions: Outcomes
and Associations1
ORIGINAL RESEARCH
Andrew D. Smith, MD
Erick M. Remer, MD
Kelly L. Cox, MD
Michael L. Lieber, MS
Brian C. Allen, MD2
Shetal N. Shah, MD
Brian R. Herts, MD
1
From the Section of Abdominal Imaging, Imaging
Institute (A.D.S., E.M.R., K.L.C., M.L.L., B.C.A., S.N.S.,
B.R.H.), Glickman Urological and Kidney Institute (E.M.R.,
B.R.H.), and Department of Quantitative Health Sciences
(M.L.L.), Cleveland Clinic Foundation, 9500 Euclid Ave,
A21, Cleveland, OH, 44195; Department of Radiology,
University of Mississippi Medical Center, Jackson, Miss
(A.D.S.). Received May 5, 2011; revision requested June
14; revision received August 5; accepted August 10; final
version accepted August 23. Address correspondence to
E.M.R. (e-mail: remere1@ccf.org).
2
Current address: Section of Abdominal Imaging, De-
partment of Radiology, Wake Forest University Baptist
Medical Center, Winston-Salem, NC.
q
RSNA, 2011
152
Purpose: To evaluate clinical outcomes, pathologic subtypes, met-
astatic disease rate, and clinical features associated with
malignancy in Bosniak category IIF and III cystic renal
lesions.
Materials and This retrospective study was institutional review board
Methods: approved and HIPAA compliant. Informed consent was
waived. Radiology and hospital information systems were
searched for Bosniak IIF and Bosniak III lesions in com-
puted tomographic (CT) reports from January 1, 1994 to
August 31, 2009. Patients 18 years and older with unen-
hanced and contrast material–enhanced CT results and
with lesions either surgically resected or with 1 year or
more of surveillance were included. Data recorded were
history of renal cell carcinoma, number of renal lesions,
presence of a coexistent solid renal mass, surgical patho-
logic findings, and presence of metastatic disease from a
renal malignancy. Sixty-two patients with 69 Bosniak IIF
lesions and 131 patients with 144 Bosniak III lesions were
identified. Proportions from independent groups were com-
pared by using the Fisher exact test; continuous variables
were compared by using a two-tailed two-sample t test or
a Wilcoxon two-sample test.
Results: The malignancy rate of resected Bosniak IIF lesions was
25% (four of 16) and that for Bosniak III lesions was 54%
(58 of 107) (P = .03). Thirteen percent (nine of 69) of Bos-
niak IIF lesions progressed at follow-up, and 50% (four of
eight) of these resected cysts were malignant. History of
primary renal malignancy, coexisting Bosniak category IV
lesion and/or solid renal mass, and multiplicity of Bosniak
III lesions were each associated with an increased malig-
nancy rate in Bosniak III lesions. No patients developed
locally advanced or metastatic disease from a Bosniak IIF
or III lesion.
Conclusion: Although the malignancy rate in surgically excised Bosniak
IIF and Bosniak III cystic renal lesions was 25% and 54%,
respectively, in our study, the malignancy rate was higher
in patients with a history of primary renal malignancy or
coexisting Bosniak IV lesion and/or solid renal neoplasm.
q
RSNA, 2011
radiology.rsna.org n Radiology: Volume 262: Number 1—January 2012
GENITOURINARY IMAGING: Bosniak Category IIF and III Lesions
T
he Bosniak renal cyst classification
system, first described in 1986 (1),
is a well-established, widely used
method that uses computed tomographic
(CT) findings to categorize cystic renal
masses into groups on the basis of im-
aging features associated with malig-
nancy. In brief, Bosniak category I and
Bosniak category II cystic renal lesions
are simple and mildly complex cysts, re-
spectively, and require no further eval-
uation. Bosniak category IV cystic renal
lesions are almost certainly malignant
and are resected in clinically able pa-
tients. Bosniak category III cystic renal
lesions are indeterminate in malignant
potential and most commonly managed
with surgical excision. While the malig-
nancy risk of Bosniak III cystic lesions is
thought to be approximately 50% (2),
reported malignancy rates range from
31% (3) to 100% (4). Furthermore, stud-
ies published to date have included rel-
atively small numbers of lesions: Four
studies described fewer than 15 Bosniak
III lesions each (3–6), one included 33
(7), and the largest included 49 (8). Bos-
niak category IIF is a modification of the
initial classification used to describe a
group of cystic lesions that are more
Advances in Knowledge
n Fifty-four percent (58 of 107) of
resected Bosniak category III
cystic renal lesions and 25%
(four of 16) of Bosniak category
IIF lesions were malignant, but
resected Bosniak IIF lesions were
highly selected and had a high
number of risk factors associated
with malignancy.
n A history of primary renal malig-
nancy, coexisting Bosniak cate-
gory IV cystic renal lesion and/or
solid renal mass (common in
patients with resected Bosniak
IIF cystic renal lesions), and mul-
tiplicity of Bosniak III cystic renal
lesions were all associated with
an increased risk of malignancy.
n No Bosniak IIF or III cystic renal
lesion progressed to a Bosniak IV
lesion or solid renal neoplasm,
and no patient developed locally
advanced or metastatic disease.
Radiology: Volume 262: Number 1—January 2012
complex than Bosniak II lesions but not
complex enough to be characterized as
Bosniak III lesions. Bosniak category
IIF cystic renal lesions generally un-
dergo imaging surveillance to demon-
strate stability over the course of time,
a feature associated with benignity (9).
There have been several studies evalu-
ating surveillance of Bosniak IIF lesions
(7,9), but because they are believed to
be benign, there are few studies that
have evaluated resected Bosniak IIF le-
sions (10).
We performed a retrospective re-
view of Bosniak IIF and Bosniak III le-
sions to evaluate the clinical outcomes,
pathologic diagnoses, metastatic disease
rate, and clinical features associated
with malignancy.
Materials and Methods
Identification of Patients with Bosniak IIF
and III Cystic Renal Lesions
Informed consent was waived in this
institutional review board–approved,
Health Insurance Portability and Ac-
countability Act–compliant retrospective
study. A database with integrated data
from our radiology (Syngo Workflow;
Siemens Medical USA, Malvern, Pa)
and hospital (Epic Systems, Verona, Wis)
information systems was searched for
the term “Bosniak” combined with “IIF”
or “2F” or “III” or “3” from January 1,
1994 to August 31, 2009 in reports
from CT or magnetic resonance (MR)
imaging studies performed in patients
18 years or older. The search of approx-
imately 100 000 reports yielded 3215
examinations. Inclusion criteria for the
study group were unenhanced and con-
trast material–enhanced CT examination
results and surgical resection of the lesion
or 1 year or more of imaging follow-up
of the lesion. Patients with Bosniak IIF
Implication for Patient Care
n These data support current rec-
ommendations for observation in
patients with Bosniak IIF or III
cystic renal lesions and multiple
comorbidities or short life
expectancy.
n radiology.rsna.org
Smith et al
or Bosniak III lesions in which biopsy
and ablation were percutaneously per-
formed and those lesions that were
not resected but had less than 1 year
of imaging follow-up were excluded from
analysis. Patients with von Hippel-Lindau
syndrome were excluded from analysis
because the great majority of cystic re-
nal neoplasms in these patients are ma-
lignant clear cell renal cell carcinoma
(RCC).
A total of 308 patients with Bosniak
IIF and/or Bosniak III lesions were iden-
tified. Of these, 113 patients were ex-
cluded from analysis because of lack of
surgical excision with less than 1 year
of imaging follow-up. Seven of these pa-
tients with eight Bosniak III lesions had
benign pathologic findings at percutane-
ous biopsy (seven core-needle biopsies
and one fine-needle aspiration biopsy)
and were excluded because of potential
sampling error. An additional two pa-
tients were excluded because of unclear
surgical pathologic findings (n = 1) or a
history of von Hippel-Lindau syndrome
(n = 1). A total of 69 Bosniak IIF le-
sions in 62 patients and 144 Bosniak III
lesions in 131 patients were included in
the final analysis.
A CT study both without contrast
material and with intravenous contrast
material was used for initial classifica-
tion with the Bosniak grading system
in 92% (178 of 193) of patients, and
MR imaging without and with intrave-
nous contrast material was used in 8%
Published online before print
10.1148/radiol.11110888 Content code:
Radiology 2012; 262:152–160
Abbreviation:
RCC = renal cell carcinoma
Author contributions:
Guarantors of integrity of entire study, A.D.S., E.M.R.;
study concepts/study design or data acquisition or data
analysis/interpretation, all authors; manuscript drafting
or manuscript revision for important intellectual content,
all authors; manuscript final version approval, all au-
thors; literature research, A.D.S., E.M.R., K.L.C., S.N.S.,
B.R.H.; clinical studies, A.D.S., E.M.R., S.N.S.; statistical
analysis, A.D.S., K.L.C., M.L.L.; and manuscript editing,
all authors
Potential conflicts of interest are listed at the end
of this article.
153
GENITOURINARY IMAGING: Bosniak Category IIF and III Lesions
(15 of 193) of patients for the initial
classification. If MR imaging was the
first examination performed, CT per-
formed later in the patient’s clinical
course was utilized for lesion character-
ization in this study.
Data Collection and Surgical Pathologic
Correlation and Follow-up
Clinical data recorded from the elec-
tronic medical record included patient
sex; date of birth; age at initial appli-
cation of Bosniak classification; date of
initial imaging examination used for
Bosniak classification; specialty of the
physician who ordered the imaging study
used to classify the lesion (urologist
or nonurologist); history of hemodi-
alysis; history of malignancy (primary
renal tumor or neoplasm originating in
another organ) or Bosniak IV lesion;
date of the last clinical note; and, when
present, the date of development of met-
astatic disease from a primary renal
malignancy. The imaging reports from
the time of initial application of the
Bosniak criteria until the last available
imaging report were assessed for each
patient, and the date of any reclassifica-
tion of a Bosniak category was recorded.
All Bosniak IIF lesions that were later
categorized as Bosniak III lesions were
categorized as Bosniak III lesions in the
pathologic and clinical analysis. Because
a number of Bosniak IIF and III lesions
were reclassified at follow-up imaging,
a radiologist with 17 years of experience
(E.M.R.) reevaluated the images to de-
termine whether reclassification resulted
from a change in the lesion imaging
characteristics or, if no change, from
differing opinions of the original inter-
preting radiologists.
In addition, the practice type of the
radiologist who initially classified the
lesion (fellowship-trained academic or
community-based radiologist), the number
and side (right and/or left kidney) of the
Bosniak IIF or III lesions, the number
and side of coexisting Bosniak IV le-
sions and/or solid renal mass, and the
type of imaging studies used to initially
classify the Bosniak cystic lesions were
recorded.
All imaging studies were reviewed to
identify the axial section thickness and to
154
determine if multiplanar reformations
were obtained and archived. For patients
with surgical resection of Bosniak IIF
and/or Bosniak III lesions, the date of
surgery was recorded, and the interval
from initial Bosniak classification to
surgery was calculated. The pathology
specimen reports from these patients
were reviewed, and the lesions were
classified as benign or malignant. The
Fuhrman nuclear grade of malignant le-
sions was recorded. For those patients
without surgical treatment, the date of
initial Bosniak classification of the cys-
tic renal lesion(s) and the date of the
last CT imaging study and last clinical
note to calculate duration of imaging
surveillance and duration of clinical sur-
veillance, respectively, were used. The
frequency of reclassification of Bosniak
IIF lesions to Bosniak III lesions as re-
ported by the interpreting radiologist
was noted. The application of Bosniak
classification by the initial interpreting
radiologist was not reevaluated.
Imaging Technique
CT imaging at our institution was ob-
tained with multidetector CT scanners
(Emotion Duo, Volume Zoom, Sensation
16, or Sensation 64; Siemens Medical,
Forchheim, Germany). Only patients who
underwent CT imaging without contrast
material and CT imaging with intrave-
nous contrast material were included;
our standard renal CT protocol included
unenhanced imaging of the kidneys, fol-
lowed by a 20-mL test bolus of contrast
material, usually iopromide (300 mg of
iodine per milliliter, Ultravist 300; Bayer
Healthcare, Berlin, Germany), for arte-
rial enhancement timing. Postcontrast
imaging was performed 5 seconds after
the optimal arterial time (vascular and
corticomedullary phase) and then again
2 minutes after the start of contrast ma-
terial injection (nephrographic phase).
The specific CT parameters, contrast ma-
terial injection rate, and contrast ma-
terial dose were not consistent in this
retrospective study. Almost all CT exam-
inations were performed with scanners
with at least four detector rows. Eight
studies were performed prior to 1999
for which scanner type could not be
determined.
radiology.rsna.org
Smith et al
Statistical Analysis
The homogeneity hypothesis for pro-
portions from independent groups was
assessed by using the Fisher exact test.
Continuous variables were compared
by using a two-tailed two-sample t test
or Wilcoxon two-sample test. A statis-
tically significant difference was consid-
ered to be indicated by a P value less
than .05.
Results
Nearly all of the imaging studies (191
[99%] of 193) were interpreted by
fellowship-trained abdominal radiolo-
gists; only two studies were interpreted
by general practice community radiolo-
gists. The section thickness of axial CT
images (170 of 178 CT examinations
available for review) was 3 mm in 81%
(137 of 170) of cases and 5 mm in 19%
(33 of 170) of cases. Most of the CT
studies (158 [82%] of 193) had mul-
tiplanar reformations available to the
interpreting radiologist at the time of
diagnosis.
Characteristics and treatment of pa-
tients with Bosniak IIF and Bosniak III
lesions are shown in Table 1. Seventy-
seven percent (101 of 131) of patients
with Bosniak III lesions underwent sur-
gical excision of the lesion(s) rather
than imaging surveillance. In contrast,
21% (13 of 62) of patients with Bosniak
IIF lesions underwent surgical excision.
Sixteen Bosniak IIF lesions in 13 pa-
tients were resected, and six patients
had a coexisting Bosniak IV lesion and/
or solid renal mass (five ipsilateral and
one contralateral to the Bosniak IIF le-
sions), three patients had a history of
primary renal malignancy, two patients
had an ipsilateral Bosniak III lesion,
three patients requested definitive sur-
gical treatment, and two patients had
recurrent cysts after aspiration and sus-
picion for malignancy by the urologist.
Patients with Bosniak IIF lesions were
significantly more likely to undergo re-
section rather than imaging surveillance
when there was a coexisting Bosniak
IV lesion and/or solid renal neoplasm
(P = .0001).
Nine cystic renal lesions, in eight pa-
tients, initially classified as Bosniak IIF
n Radiology: Volume 262: Number 1—January 2012
GENITOURINARY IMAGING: Bosniak Category IIF and III Lesions Smith et al

Table 1
Characteristics and Treatment of Patients with Bosniak IIF and III Lesions
Bosniak IIF Lesions Bosniak III Lesions
Parameter Surgical Excision Imaging Surveillance P Value Surgical Excision Imaging Surveillance P Value

No. of patients 13 49 … 101 30 …

No. of Bosniak IIF or III lesions 16 53 … 107 37 …
Median age (y)* 52 (33–74) 66 (37–87) .0043† 57 (27–80) 68 (43–89) .00007†
Percentage of imaging studies ordered by a 100 (13/13) 57 (28/49) .0026‡ 90 (91/101) 63 (19/30) .0012‡
urologist
Percentage of patients with history of 0 (0/13) 20 (10/49) .10 3 (3/101) 27 (8/30) .0003‡
nonrenal metastatic neoplasm
Percentage of patients with coexisting Bosniak 46 (6/13) 0 (0/49) .0001‡ 16 (16/101) 23 (7/30) .41
IV lesion and/or solid renal mass
Percentage of patients with history of 23 (3/13) 4 (2/49) .06§ 7 (7/101) 30 (9/30) .0021‡
malignant primary renal neoplasm
Median duration of imaging surveillance after … 2.4 (1.0–8.0) … … 3.8 (1.0–8.2) …
classification (y)*
Median duration of clinical surveillance after 3.1 (0.1–7.4) 3.6 (1.0–8.5) … 2.1 (0–11.1) 4.4 (1.0–15.2) …
classification (y)*
Percentage of patients with metastatic disease 0 (0/13) 0 (0/49) … 0 (0/101) 0 (0/30) …
from a primary renal malignancy

* Data in parentheses are ranges.

†
Statistically significant P value (,.05, two-sample t test).
‡
Statistically significant P value (,.05, Fisher exact test).
§
Marginally statistically significant P value (.05 , P , .10, Fisher exact test).

lesions (nine [13%] of 69) were later
reclassified as Bosniak III lesions at a
median time of 0.5 years (range, 0.2–
2.0 years) after the initial classification
as a Bosniak IIF lesion (Fig 1). A single
radiologist (E.M.R.) reviewed these pa-
tients’ studies and determined that mor-
phologic changes led to reclassification
of three lesions (Fig 2), Bosniak III was
the appropriate original classification
in four lesions (Fig 3), and Bosniak IIF
was the appropriate classification in two
lesions such that the lesion should not
have been reclassified. Eight of these le-
sions were resected with a malignancy
rate of 50% (four of eight). One patient
had a history of contralateral partial
nephrectomy for papillary RCC, was
treated with periodic surveillance imag-
ing for 1.9 years, and subsequently was
lost to follow-up.
After initial CT categorization, pa-
tients with Bosniak IIF lesions under-
went surgery at a median of 0.1 year
(range, 0–6.1 years), and those with
Bosniak III lesions underwent surgery
at a median of 0.1 year (range, 0–2.1
years). Patients with either Bosniak IIF
or Bosniak III lesions who underwent
imaging surveillance were on average
significantly older than those patients
who underwent surgical excision, with a
mean difference in both cases of greater
than 10 years (Table 1). Clinical history
that may have affected the decision to
perform surgical resection such as co-
existing renal lesions, history of nonre-
nal malignancy, and ordering physician
are listed in Table 1.
The malignancy rate of and specific
pathologic subtypes of resected Bosniak
IIF and Bosniak III lesions are shown in
Table 2. Fifty-four percent (58 of 107)
of resected Bosniak III lesions were
malignant and 25% (four of 16) of re-
sected Bosniak IIF lesions were malig-
nant, which was a significant difference
(P = .03, Fisher exact test; estimated
odds ratio = 3.55; 95% confidence in-
terval: 1.08, 11.72). The Fuhrman nu-
clear grade was only available for three
of four resected malignant Bosniak IIF
lesions and was grade 1 for one lesion
and grade 2 for two lesions. Fuhrman
nuclear grade was available for 57 of
the 58 resected malignant Bosniak III
lesions and was grade 1 in 19% (11 of
57), grade 2 in 61% (35 of 57), grade 3
in 19% (11 of 57), and grade 4 in 0% (0
of 57) of lesions. Clear cell carcinoma
was the most prevalent malignant neo-
plasm (28 [48%] of 58), and the major-
ity of benign lesions (26 [53%] of 49)
were epithelial or fibrous cysts.
The association of three factors and
the prevalence of malignancy in Bosniak
IIF and Bosniak III lesions were evalu-
ated: coexisting Bosniak III lesion, co-
existing Bosniak IV lesion and/or solid
mass, and a history of malignant renal
neoplasm (Table 3). Two patients had
both Bosniak IIF and Bosniak III lesions.
The first patient had ipsilateral solitary
Bosniak IIF and Bosniak III lesions with
a benign cortical cyst and cystic nephro-
ma, respectively, at surgical pathologic
findings. The second had ipsilateral
solitary Bosniak IIF and Bosniak III
lesions and two contralateral Bosniak
IIF lesions that were all benign cortical
cysts according to surgical pathologic

Radiology: Volume 262: Number 1—January 2012 n radiology.rsna.org 155

GENITOURINARY IMAGING: Bosniak Category IIF and III Lesions Smith et al

Figure 1 a coexisting Bosniak IV lesion at the

time of Bosniak classification.
The duration of imaging and clinical
surveillance of Bosniak IIF and Bosniak
III lesions included in the study is pre-
sented in Table 1. The majority of pa-
tients with resected Bosniak IIF (nine
[69%] of 13) and Bosniak III lesions
(88 [87%] of 101) had no imaging fol-
low-up after surgery. Despite a higher
percentage of patients with a coexist-
ing Bosniak IV lesion and/or solid renal
neoplasm or history of malignant pri-
mary renal neoplasm in patients with
Bosniak III lesions being treated with
imaging surveillance of 1 year or longer
compared with surgical excision, no
patient in this study developed locally
advanced or metastatic disease from
a cystic renal neoplasm. No Bosniak
III lesions (n = 144) progressed to a
Bosniak IV lesion; seven were down-
graded to Bosniak IIF lesions. Median
imaging follow-up was 3.8 years (range,
1.0–8.2 years), and median clinical
follow-up was 4.4 years (range, 1.0–
15.2 years) for patients with Bosniak
III lesions managed by using imaging
surveillance (Table 1).

Discussion
The Bosniak classification system is
Figure 1: Transverse baseline (a) unenhanced and (b) nephrographic phase CT images in 69-year-old widely acknowledged as a useful guide-
man with a cystic renal lesion show thick partially calcified wall (arrow) with perceived enhancement inter- line for managing cystic renal masses
preted as Bosniak IIF lesion. Transverse (c) unenhanced and (d) nephrographic phase CT images at 6-month (11). The malignancy rate of 54% (58
follow-up show interval development of ill-defined enhancement along the posterior wall (arrow) leading to of 107) in our patient population is
reclassification as Bosniak III lesion. Pathologic diagnosis was benign cyst with fibrous wall. consistent with the previously published
and widely held belief that approxi-
findings. For Bosniak IIF lesions, sig- patients with benign Bosniak IIF lesions mately 50% of Bosniak III lesions are
nificantly more lesions were malignant (n = 9), the median age was 48 years malignant. While previous studies have
when the patient had a coexisting (range, 33–74 years) (P = .13, Wilcoxon shown a surprising variability in the
Bosniak IV lesion and/or a solid mass. two-sample test). The median age of pa- percentage of malignant Bosniak III
For Bosniak III lesions, any of the three tients with malignant Bosniak III lesions cysts (31%–100%), these numbers were
factors was associated with a signifi- (n = 55) was 58 years (range, 27–80 based on small series that used older
cantly higher percentage of malignant years), and for patients with benign CT technology. A weighted average from
lesions. Two of these factors were more Bosniak III lesions (n = 46), the median six studies that each included more
common in patients with Bosniak III le- age was 57 years (range, 33–78 years) than 30 renal cystic masses found 65%
sions treated with imaging surveillance (P = .38, Wilcoxon two-sample t test). of Bosniak III lesions to be malignant
than resection (Table 1). A history of hemodialysis was un- (12). As far as we are aware, this series
Patient age was not significantly dif- common in our patient population. One of 107 resected Bosniak III lesions
ferent between patients with Bosniak patient with a single resected benign represents the largest report of path-
IIF and those with Bosniak III lesions. Bosniak III lesion was receiving hemo- ologically examined Bosniak III lesions
The median age of patients with ma- dialysis, and one patient with two re- imaged by using multiphasic multide-
lignant Bosniak IIF lesions (n = 4) was sected malignant Bosniak IIF lesions tector CT and multiphasic MR imaging
62 years (range, 50–71 years), and for was receiving hemodialysis but also had examinations.

156 radiology.rsna.org n Radiology: Volume 262: Number 1—January 2012

GENITOURINARY IMAGING: Bosniak Category IIF and III Lesions
Figure 2
Figure 2: Transverse baseline (a) unenhanced and (b) nephrographic phase CT images in 61-year-old
woman with a cystic renal lesion show a multicystic lesion with multiple septa of varying thickness (arrow),
and lesion was interpreted as Bosniak IIF. Transverse (c) unenhanced and (d) nephrographic phase CT
images at 15-month follow-up show progressive septal thickening (arrow) that was categorized as Bosniak
III. Pathologic diagnosis was clear cell carcinoma.
The malignancy rate of 54% in our
series differs from the malignancy rate
of 82% in 33 resected Bosniak III cysts
found by O’Malley et al (7). The high
malignancy rate was attributed to the
addition of the Bosniak IIF category be-
tween the publication of older studies
and the analysis. The use of the Bosniak
IIF category was thought to increase
the accuracy of the classification system
by downgrading lesions that were not
complex enough to be characterized as
category III but that were more complex
than category II lesions (9). There are
several explanations why we may not
have found as high a malignancy rate
in our Bosniak III group. Because the
Radiology: Volume 262: Number 1—January 2012 n radiology.rsna.org
Bosniak criteria were developed at the
institution of O’Malley et al, the IIF
category was more likely to have been
immediately adopted, universally used,
and understood. Furthermore, images
in the O’Malley series were read by
one of several faculty radiologists. This
suggests that a small group may have
evaluated these studies, leading to less
observer variability. Our images were
interpreted by a group of 12 fellowship-
trained abdominal radiologists. While
they were all well acquainted with the
Bosniak classification, it is unclear how
often or consistently classifications were
applied because the search parameters
for this study only captured lesions that
Smith et al
were classified by using the Bosniak
system. Thirteen percent of Bosniak IIF
lesions were reclassified as Bosniak III
lesions at surveillance imaging in our
series, which is similar to the 14.8% rate
in the O’Malley et al series ( 7 ), but
more than the 4.8% in 48 cysts in the
Israel and Bosniak series (9). Fifty per-
cent of those that were resected were
malignant. On the basis of image review,
reclassification in the series was more
commonly because of reader disagree-
ment than a morphologic change in the
lesion.
There was a surprisingly high ma-
lignancy rate of 25% (four of 16) in re-
sected Bosniak IIF lesions. However, the
number of resected Bosniak IIF lesions
was small and only reflected a subset of
Bosniak IIF lesions in the population.
The patient population with resected
Bosniak IIF lesions was highly selected
and had a higher number of risk fac-
tors associated with malignancy than
the imaging surveillance group. In our
series, in patients with a Bosniak IIF le-
sion and a coexistent Bosniak IV lesion
and/or solid mass, 71% (five of seven)
of the Bosniak IIF lesions were malig-
nant. Similarly, 67% (two of three) of
Bosniak IIF lesions in patients with a
history of RCC were malignant. These
correlations certainly suggest that there
are unaccounted for variables that could
have affected the malignancy rate in
resected Bosniak IIF lesions in this
series.
Bosniak IIF and Bosniak III lesions
encompassed a wide variety of neo-
plasms. While the most common malig-
nant renal tumor in the series was clear
cell carcinoma, it represented less than
50% of the malignant cystic lesions. A
cystic growth pattern is seen in only
4%–15% of all RCCs (13) and occurs
by four mechanisms: intrinsic unilocu-
lar cystic growth, intrinsic multilocular
cystic growth, origin in the epithelium
of a simple cyst, or cystic necrosis (14).
True cystic growth, rather than necro-
sis, appears to confer a better prognosis
(15). Several prior studies of cysts with
Bosniak grading have concentrated on
the distinction between malignant and
benign lesions and have not reported
the subtypes of carcinoma (3,6,8). The
157
GENITOURINARY IMAGING: Bosniak Category IIF and III Lesions Smith et al

Figure 3 if near water attenuation was measured

on precontrast CT images.
Clinical data and CT image findings
strongly affected the risk of malignancy
for complex cysts in this study. Seventy-
one percent (five of seven) of Bosniak
IIF lesions found in patients with a co-
existing Bosniak IV lesion and/or solid
renal mass were malignant, which oc-
curred significantly more often than if
either of these findings was not present.
The percentage of Bosniak III lesions
that were malignant was especially high
when associated with other Bosniak III
lesions in the same patient (10 [83%]
of 12), a coexisting Bosniak IV lesion
and/or solid renal mass (15 [83%] of
18), or a history of RCC (eight [100%]
of eight). This suggests that patient-
related factors that predispose to RCC
may be at work in these patients. As
with all Bosniak III cystic renal lesions,
surgery should be first-line management.
If there are contraindications to sur-
gery, imaging follow-up is a reasonable
approach.
None of the patients with Bosniak
IIF or Bosniak III lesions in this study
(n = 193) had progression to Bosniak
IV lesions or solid neoplasms, and none
developed locally advanced or meta-
static disease. If there is a continuous
spectrum such that some Bosniak III
Figure 3: Transverse baseline (a) unenhanced and (b) nephrographic phase CT images in 59-year-old lesions progress to Bosniak IV lesions,
woman with a cystic renal lesion show a thick wall (arrow in b), with linear and punctate calcifications the rate of progression was too low to
(arrows in a and c) that measured 30 HU on unenhanced images but did not enhance, and lesion was be captured in our series. Despite a
interpreted as Bosniak IIF. Transverse (c) unenhanced and (d) nephrographic phase CT images at 6-month higher number of potential risk factors
follow-up show no change, but the lesion was interpreted as Bosniak III, presumably because of measurable associated with malignant Bosniak III le-
wall enhancement (arrow in d). Pathologic diagnosis was benign cyst. sions in the imaging surveillance group
(n = 30), none of these consequences
presumption that a malignant cystic le- papillary RCC is hyperdense and ho- were detected with imaging follow-up
sion equates to a cystic clear cell car- mogeneously enhancing at CT (16). A for a median of 3.8 years (range, 1.0–
cinoma and that a benign cystic lesion papillary RCC can, however, occasion- 8.2 years) and clinical follow-up for a
equates to an epithelial cyst is false. In ally manifest as a cystic mass. This has median of 4.4 years (range, 1.0–15.2
truth, there are a host of benign and been attributed to inherent architecture years).
malignant cystic renal neoplasms, in- or secondary to cystic degeneration Although the rate of development of
cluding cystic clear cell carcinoma, mul- and extensive necrosis (17). One series metastatic disease found at short-term
tilocular cystic RCC, cystic nephroma, found that cystic degeneration was al- imaging surveillance for Bosniak IIF and
and mixed epithelial and stromal tumor most as likely to occur in papillary RCC Bosniak III lesions was 0% (0 of 193),
of the kidney. Each confers a different as in clear cell RCC (18). In addition, the long-term rate of development of
prognosis and potentially requires dif- some papillary RCC are known to en- metastatic disease remains unknown.
ferent therapy. hance only minimally after intravenous Despite a malignancy rate of more than
We were surprised to find that 25% contrast material administration. These 50% for Bosniak III lesions, there is no
(15 of 58) of malignant resected cys- minimally enhancing tumors may have prospective data to show that resection
tic lesions were papillary RCCs at his- been mistaken for cystic lesions at the improves survival. Given that nearly
topathologic examination. Typically, time of image interpretation, especially half of Bosniak III lesions are benign

158 radiology.rsna.org n Radiology: Volume 262: Number 1—January 2012

GENITOURINARY IMAGING: Bosniak Category IIF and III Lesions Smith et al

Table 2 vers evaluating the features that are used

Surgical Pathologic Findings in Excised Bosniak IIF and III Lesions to select a Bosniak category (20) and
agreement between category assign-
All Excised Bosniak IIF Lesions All Excised Bosniak III Lesions ments (6) have been previously eval-
Surgical Pathologic Finding (n = 16) (n = 107)
uated. A choice was made not to rein-
Malignant 25 (4/16)* 54 (58/107)* terpret studies by using expert readers,
Clear cell RCC 19 (3/16) 26 (28/107) with the hope that this series would
Papillary RCC … 14 (15/107) represent a real-world experience, not
Unclassified RCC … 5 (5/107) an idealized one. Third, we have no way
Mixed RCC 6 (1/16) 4 (4/107) of knowing how often or consistently
Multilocular cystic RCC … 5 (6/107) Bosniak criteria were applied to all cystic
Benign 75 (12/16) 46 (49/107) renal lesions encountered at CT imaging,
Benign epithelial or fibrous cyst 31 (5/16) 24 (26/107) because the search parameters for this
Cystic nephroma 25 (4/16) 12 (13/107) study only captured lesions that were
Mixed epithelial stromal tumor … 1 (1/107) classified by using the Bosniak system.
Oncocytoma … 1 (1/107) Fourth, for patients with resected
Other benign tissue or few atypical cells 19 (3/16) 8 (8/107)
Bosniak IIF lesions, the statistical power
Note.—Data are percentages, with numbers used to calculate percentages in parentheses. Percentages may not add up to associate malignant Bosniak IIF le-
because of rounding. sions with risk factors such as coexist-
* The malignancy rate of resected Bosniak IIF and III lesions was significantly different (P = .03, Fisher exact test). ing Bosniak III lesion or history of RCC
was limited by patient sample size, and
the lack of a significant P value does
not include or exclude an important
Table 3
clinical association (Table 3). Fifth, in
Risk Factors for Malignancy in Resected Bosniak IIF and III Lesions lesion-level analyses (Table 2), multiple
lesions within patients were treated
Prevalence of Malignant in Resected Prevalence of Malignant in Resected
Risk Factor Bosniak IIF Lesions (n = 16) Bosniak III Lesions (n = 107)
as statistically independent; therefore,
any intrapatient correlation due to clus-
Coexisting Bosniak III lesion tering was not accounted for in hypo-
Present 0 (0/2) 83 (10/12) thesis tests. We note that 23% (three
Absent 43 (6/14) 51 (48/95) of 13) of patients with surgically excised
P value .50 .036* Bosniak IIF lesions and 8% (eight of
Coexisting Bosniak IV lesion 101) of patients with surgically excised
and/or solid renal mass
Bosniak III lesions had more than one
Present 71 (5/7) 83 (15/18)
Bosniak IIF or Bosniak III lesion, and only
Absent 11 (1/9) 49 (44/89)
one patient had more than three. Last,
P value .035* .009*
the true malignancy rate of the Bosniak
History of malignant primary
IIF and Bosniak III lesions in patients
renal cell neoplasm
Present 67 (2/3) 100 (8/8)
who underwent imaging surveillance
Absent 31 (4/13) 45 (44/99) is not known, and the length of imag-
P value .52 .0067* ing follow-up may have been insuffi-
cient to detect imaging manifestations of
Note.—Unless otherwise indicated, data are percentages of patients, with numbers used to calculate percentages in malignancy.
parentheses.
In conclusion, 54% (58 of 107) of
* Significant P value (,.05) by using the Fisher exact test.
Bosniak III cystic renal lesions were ma-
lignant in our series, with clear cell car-
cinoma and papillary RCC most com-
and that the risk of metastatic disease case series; therefore, there was no mon. Thirteen percent (nine of 69) of
or local progression appears to be low, standardized follow-up interval and there Bosniak IIF lesions were reclassified at
our data support current recommenda- were changes in CT technology during follow-up and 50% (four of eight) of
tions for observation in patients with the time interval of the study. Second, these patients had malignant lesions.
Bosniak IIF or Bosniak III lesions and there were multiple readers who in- A history of primary renal malignancy,
multiple comorbidities or short life ex- terpreted the CT images that were in- coexisting Bosniak IV cystic renal lesion
pectancy (19). cluded in the study. Application of the and/or solid renal mass (common in
There were a number of limitations of Bosniak classification system is subjec- patients with resected Bosniak IIF cys-
our study. First, it was a retrospective tive, and the agreement between obser- tic renal lesions), and multiplicity of

Radiology: Volume 262: Number 1—January 2012 n radiology.rsna.org 159

GENITOURINARY IMAGING: Bosniak Category IIF and III Lesions
Bosniak III cystic renal lesions were all
associated with an increased risk of
malignancy in Bosniak IIF and Bosniak
III cystic renal lesions. None of the pa-
tients with Bosniak IIF or III cystic renal
lesions in this study had progression to
Bosniak IV lesions or solid renal neo-
plasms, and none developed locally ad-
vanced or metastatic disease.
Disclosures of Potential Conflicts of Inter-
est: A.D.S. No potential conflicts of interest
to disclose. E.M.R. No potential conflicts of
interest to disclose. K.L.C. No potential con-
flicts of interest to disclose. M.L.L. No po-
tential conflicts of interest to disclose. B.C.A.
No potential conflicts of interest to disclose.
S.N.S. No potential conflicts of interest to
disclose. B.R.H. Financial activities related to
the present article: none to disclose. Financial
activities not related to the present article: in-
stitution has grant from Siemens Medical for
radiation dose research. Other relationships:
none to disclose.
References
1. Bosniak MA. The current radiological ap-
proach to renal cysts. Radiology 1986;
158(1):1–10.
2. Smith ND, Campbell SC. Evaluation of the re-
nal mass. In: Kursh ED, Ulchaker JC, eds. Of-
fice urology: the clinician’s guide. Totowa, NJ:
Humana, 2001; 145–163.
3. Cloix P, Martin X, Pangaud C, et al. Surgical
management of complex renal cysts: a series
of 32 cases. J Urol 1996;156(1):28–30.
4. Wilson TE, Doelle EA, Cohan RH, Wojno K,
Korobkin M. Cystic renal masses: a re-
160
evaluation of the usefulness of the Bosniak
classification system. Acad Radiol 1996;
3(7):564–570.
5. Aronson S, Frazier HA, Baluch JD,
Hartman DS, Christenson PJ. Cystic renal
masses: usefulness of the Bosniak classification.
Urol Radiol 1991;13(2):83–90.
6. Siegel CL, McFarland EG, Brink JA,
Fisher AJ, Humphrey P, Heiken JP. CT of
cystic renal masses: analysis of diagnostic
performance and interobserver varia-
tion. AJR Am J Roentgenol 1997;169(3):
813–818.
7. O’Malley RL, Godoy G, Hecht EM,
Stifelman MD, Taneja SS. Bosniak category
IIF designation and surgery for complex re-
nal cysts. J Urol 2009;182(3):1091–1095.
8. Curry NS, Cochran ST, Bissada NK. Cystic
renal masses: accurate Bosniak classification
requires adequate renal CT. AJR Am J
Roentgenol 2000;175(2):339–342.
9. Israel GM, Bosniak MA. Follow-up CT of
moderately complex cystic lesions of the
kidney (Bosniak category IIF). AJR Am J
Roentgenol 2003;181(3):627–633.
10. Spaliviero M, Herts BR, Magi-Galluzzi C,
et al. Laparoscopic partial nephrectomy for
cystic masses. J Urol 2005;174(2):614–619.
11. Warren KS, McFarlane J. The Bosniak
classification of renal cystic masses. BJU Int
2005;95(7):939–942.
12. Graumann O, Osther SS, Osther PJ.
Characterization of complex renal cysts: a
critical evaluation of the Bosniak classifica-
tion. Scand J Urol Nephrol 2011;45(2):
84–90.
radiology.rsna.org
Smith et al
13. Hartman DS, Davis CJ Jr, Johns T,
Goldman SM. Cystic renal cell carcinoma.
Urology 1986;28(2):145–153.
14. Eble JN, Sauter G, Epstein JI, Sesterhenn IA,
eds. Pathology and genetics of tumours of
the urinary system and male genital organs.
Lyon, France: IARC Press, 2004.
15. Han KR, Janzen NK, McWhorter VC, et al.
Cystic renal cell carcinoma: biology and
clinical behavior. Urol Oncol 2004;22(5):
410–414.
16. Silverman SG, Mortele KJ, Tuncali K,
Jinzaki M, Cibas ES. Hyperattenuating renal
masses: etiologies, pathogenesis, and imag-
ing evaluation. RadioGraphics 2007;27(4):
1131–1143.
17. Vikram R, Ng CS, Tamboli P, et al. Papillary
renal cell carcinoma: radiologic-pathologic
correlation and spectrum of disease. Radio-
Graphics 2009;29(3):741–754.
18. Brinker DA, Amin MB, de Peralta-Venturina M,
Reuter V, Chan DY, Epstein JI. Exten-
sively necrotic cystic renal cell carcinoma:
a clinicopathologic study with comparison
to other cystic and necrotic renal cancers.
Am J Surg Pathol 2000;24(7):988–995.
19. Berland LL, Silverman SG, Gore RM, et al.
Managing incidental findings on abdomi-
nal CT: white paper of the ACR incidental
findings committee. J Am Coll Radiol 2010;
7(10):754–773.
20. Benjaminov O, Atri M, O’Malley M, Lobo K,
Tomlinson G. Enhancing component on CT
to predict malignancy in cystic renal masses
and interobserver agreement of different
CT features. AJR Am J Roentgenol 2006;
186(3):665–672.
n Radiology: Volume 262: Number 1—January 2012